metadata
license: apache-2.0
tags:
- automatic-speech-recognition
- mozilla-foundation/common_voice_8_0
- generated_from_trainer
- robust-speech-event
datasets:
- common_voice
model-index:
- name: xls-r-300m-dv
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Common Voice 8
type: mozilla-foundation/common_voice_8_0
args: dv
metrics:
- name: Test WER
type: wer
value: null
- name: Test CER
type: cer
value: null
xls-r-300m-dv
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - dv dataset. It achieves the following results on the evaluation set:
- Loss: 0.6182
- Wer: 0.5481
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0003
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 4.3529 | 2.63 | 400 | 1.2050 | 0.9526 |
| 0.7191 | 5.26 | 800 | 0.6037 | 0.7216 |
| 0.3981 | 7.89 | 1200 | 0.5048 | 0.6225 |
| 0.2888 | 10.52 | 1600 | 0.5345 | 0.6170 |
| 0.2229 | 13.16 | 2000 | 0.5261 | 0.6015 |
| 0.1865 | 15.79 | 2400 | 0.5983 | 0.5924 |
| 0.1542 | 18.42 | 2800 | 0.5900 | 0.5770 |
| 0.1401 | 21.05 | 3200 | 0.6425 | 0.5783 |
| 0.1205 | 23.68 | 3600 | 0.6322 | 0.5760 |
| 0.1105 | 26.31 | 4000 | 0.6302 | 0.5567 |
| 0.0958 | 28.94 | 4400 | 0.6182 | 0.5481 |
Framework versions
- Transformers 4.16.0.dev0
- Pytorch 1.10.1+cu102
- Datasets 1.17.1.dev0
- Tokenizers 0.11.0